The 12 audit checks
Every GEO audit runs these 12 checks across three pillars. Each check produces a score, a severity rating (pass, warning, or fail), and a recommendation.
Technical AI-readiness
Can AI crawlers access your content?
AI Crawler Access
Checks whether your robots.txt allows or blocks the major AI crawlers: GPTBot, ClaudeBot, Google-Extended, and others.
Schema Markup
Looks for JSON-LD structured data: Organization, Article, FAQ, Breadcrumb, and date properties. Checks whether required properties are present on each type.
LLMs.txt
Checks for a /llms.txt file that tells AI models which pages to prioritize. A newer standard, but adoption is growing.
Content Freshness
Measures how recently the page was updated using HTTP headers and meta tags. Stale content gets cited less.
JavaScript Rendering Gap
Compares the raw HTML to the fully rendered page. If most of your content loads via JavaScript, AI crawlers may not see it.
Structural extractability
Can AI extract and cite your content?
Heading Hierarchy
Checks for a single H1, sequential heading levels (no skips from H2 to H4), and descriptive heading text that AI can use as section labels.
Answer First
Checks whether sections lead with the key takeaway. AI models prefer content that puts the answer before the explanation.
Fact Density
Counts verifiable facts per 100 words. Higher density gives AI more citable material to work with.
Passage Self-Containment
Checks whether each paragraph makes sense on its own, without needing the surrounding text. AI often extracts single passages.
Markdown Fidelity
Converts your page to markdown and measures how much content survives. Deeply nested layouts and complex CSS can strip away useful text.
Entity strength
Does AI know your brand correctly?
Knowledge Graph
Searches Wikidata for your brand entity and checks how many properties are filled in. A complete knowledge graph entry helps AI systems identify you.
AI Brand Check
Asks multiple AI providers about your brand and checks whether they describe it accurately. Inconsistent or wrong answers mean your brand signals need work.